Estimation of potential evapotranspiration of a coastal savannah environment; comparison of methods

Six potential evapotranspiration models namely, Penman-Monteith, Hargreaves-Samani , Priestley-Taylor, IRMAK1, IRMAK2 and TURC, were used to estimate daily PET values at Atomic-Kwabenya in the coastal savannah environment of Ghana for the year 2005. The study compared PET values generated by six models and identified which ones compared favourably with the Penman-Monteith model which is the recommended standard method for estimating PET. Cross comparison analysis showed that only the daily estimates of PET of Hargreaves-Samani model correlated reasonably (r = 0.82) with estimates by the Penman-Monteith model. Additionally, PET values by the Priestley-Taylor and TURC models were highly correlated (r = 0.99) as well as those generated by IRMAK2 and TURC models (r = 0.96). Statistical analysis, based on pair comparison of means, showed that daily PET estimates of the Penman-Monteith model were not different from the Priestley-Taylor model for the Kwabenya-Atomic area located in the coastal savannah environment of Ghana. The Priestley-Taylor model can be used, in place of the Penman-Monteith model, to estimate daily PET for the Atomic-Kwabenya area of the coastal savannah environment of Ghana. The Hargreaves-Samani model can also be used to estimate PET for the study area because its PET estimates correlated reasonably with those of the Penman-Monteith model (r = 0.82) and requires only air temperature measurements as inputs. (au)